Tools
Tools
Murali, Sripriya Kumari (2025) A Comparative Study Of Custom CNN And MobileNetv2 For Binary Waste Classification Using Deep Learning. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (830kB) | Preview |
Preview |
PDF (Configuration Manual)
Download (752kB) | Preview |
Abstract
The onset of significant environmental impact due to poor disposal of waste requires a new way that can facilitate waste classification in an efficient manner. This research paper looks into how deep learning models can be used in performing a binary classification of waste types through image-based information, to predict whether the waste is Organic or Recyclable waste. The main goal was to compare the performance of a Custom Convoluted Neural Network (CNN) to that of a pre-trained MobileNetV2 on the basis of the performance accuracy of the classification model in the dataset and the generalization abilities. It has a systematic approach that entails data preprocessing, design of the model, training, and testing. The two models were trained and evaluated based on the idea of a curated dataset, and the performance of the models was determined using various parameters like accuracy, precision, recall, and F1-score.
The Custom CNN model tested better with an accuracy of 92.24% compared to the MobileNetV2, which achieved 90.03% accuracy. The classification report indicated the excellence of Custom CNN in recalling the recyclable waste, but MobileNetV2 was consistent in recognizing both classes thematically. The results show that custom CNN architectures can perform better in the particular waste classification task, but MobileNetV2 is also a potentially attractive option to use in lightweight deployment settings. The study highlights the applied knowledge of deep learning in facilitating the implementation of intelligent waste management systems and more environmental sustainability. Future developments will involve multi-class classification, applying the model to edge devices, and into actual municipal waste handling systems.
| Item Type: | Thesis (Masters) |
|---|---|
| Supervisors: | Name Email Horn, Christian UNSPECIFIED |
| Subjects: | G Geography. Anthropology. Recreation > GE Environmental Sciences T Technology > TD Environmental technology. Sanitary engineering Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning |
| Divisions: | School of Computing > Master of Science in Data Analytics |
| Depositing User: | Ciara O'Brien |
| Date Deposited: | 26 Aug 2026 09:30 |
| Last Modified: | 26 Aug 2026 09:30 |
| URI: | https://norma.ncirl.ie/id/eprint/9648 |
Actions (login required)
 |
View Item |